. 2011 May 20;89(7):521–527. doi: 10.2471/BLT.10.085530

Table 1. Incidence rate ratios (IRR)^a for predictors of diarrhoea in children aged under 5 years derived using a random effects negative binomial model,^b Matlab, Bangladesh, 2005–2006.

Predictor	IRR	95% CI	P
Deep tube well use	0.541	0.308–0.949	0.032
No. of children aged < 5 years in the household	1.988	1.434–2.755	0.0001
Maternal educational level	0.932	0.858–1.012	0.094
Household socioeconomic status	0.983	0.840–1.151	0.835
Latrine density^c	0.977	0.882–1.082	0.650
Population density in the vicinity of the household^d	0.999	0.987–1.012	0.911
Study year^e	1.371	0.914–2.058	0.128

CI, confidence interval.

^a An IRR of 1 indicates that the independent variable has no effect on the risk of diarrhoea, while an IRR greater than 1 indicates an increased risk and an IRR below 1, a decreased risk.

^b The model considered 927 observations from a total of 543 households.

^c Number of latrines within 30 metres of a household.

^d Number of people living within 30 metres of a household.

^e The study years were 2005 and 2006.

Table 1. Incidence rate ratios (IRR)a for predictors of diarrhoea in children aged under 5 years derived using a random effects negative binomial model,b Matlab, Bangladesh, 2005–2006.

Table 1. Incidence rate ratios (IRR)^a for predictors of diarrhoea in children aged under 5 years derived using a random effects negative binomial model,^b Matlab, Bangladesh, 2005–2006.